[
https://issues.apache.org/jira/browse/CASSANDRA-15854?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17201439#comment-17201439
]
Marcus Eriksson edited comment on CASSANDRA-15854 at 9/24/20, 11:09 AM:
------------------------------------------------------------------------
Rewrote the test again, ran it in a loop 80 times with lower resources, no
failures
Noticed that we stream even if the repair fails right before we get the
validation response, that might have been what caused your failure.
cci run:
https://app.circleci.com/pipelines/github/krummas/cassandra/541/workflows/8d3cbf9c-6556-45d9-b993-e258b3211323
seems {{IncrementalRepairCoordinatorTimeoutTest}} failed, looks unrelated and
passes locally
was (Author: krummas):
Rewrote the test again, ran it in a loop 80 times with lower resources, no
failures
Noticed that we stream even if the repair fails right before we get the
validation response, that might have been what caused your failure.
> Truncation should fail any ongoing repairs
> ------------------------------------------
>
> Key: CASSANDRA-15854
> URL: https://issues.apache.org/jira/browse/CASSANDRA-15854
> Project: Cassandra
> Issue Type: Bug
> Components: Consistency/Repair
> Reporter: Marcus Eriksson
> Assignee: Marcus Eriksson
> Priority: Normal
> Fix For: 4.0-beta
>
>
> Truncation may race with ongoing repairs, making it possible to clear data on
> one node but then stream data its truncation would have deleted from another
> node. We should abort any ongoing preview repairs if we get a truncation
> request.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
